Best practices for internal quality control in the clinical laboratory

In the clinical laboratory, robust internal quality control (IQC) practices are essential for providing results that are consistent and fit for purpose. The delivery of quality patient care relies on precise and accurate results, and a poor IQC approach can lead to incorrect patient results and poor patient care.

In this series of three SelectScience^® webinars, Prof. Astrid Petersmann, Head of Institute of Clinical Chemistry and Laboratory Medicine at the University of Oldenburg, and Ger Kennedy, VP Informatics at Technopath Clinical Diagnostics, discuss IQC best practices and how to minimize error in datasets, as well as sharing tips and tricks on how to increase the quality of your results.

Read on for the live Q&A session from the first installment in the series, or register to watch the webinars at a time that suits you.

How many times a day should you run QC?

AP: Let's say, for glucose, the minimum requirements in my country are to do two levels on each instrument each day. That means you have two QC measurements for the whole day on an instrument where you might release hundreds of results.

It may sound scary, but the instruments and the assays are so stable that I'm actually fine with that frequency. However, there are other assays that are not that stable, or there may be issues with an instrument, or I may know that my main supplier of reagents has an issue with a certain assay. In these cases, I would increase the minimum requirements by double or even more than that. I know that is not a precise answer, but that's exactly what lab work is about. You need to get a feeling of what is helpful and work out what is a good compromise that does not jeopardize patient care.

Should the IQC target be locally derived? If so, which protocol is best?

AP: That depends on three main factors: where you're operating, in which system, and what the local regulations are. If you derive it locally in your lab, you might want to stick to rules like Westgard and run a pre-period to see whether it is stable, and then come up with your target value. It would be helpful to perform it on more than one instrument and in more than one lab. There are concepts where at least three labs with the same kind of instruments contribute to coming up with the target value for an IQC. It depends on what the regulations in that nation are.

Should you assign your own target values that are different to the manufacturer's? Or is it advisable to look at a chart that shows a manufacturer's mean and range and also your own chart with the calculated means and ranges?

AP: Usually, the manufacturers will put a lot of effort into assigning the target values. If you don't hit the target value of the manufacturer, it's worth asking yourself, what's the reason?

If you have an instrument that not very many other customers have, the target values sometimes are specific for instruments. In this case, I would advise seeing whether you can find another laboratory with the same device to back up your decision in coming up with your own target values. If your national regulations allow for it, that is a possible solution. But I would allow for a long enough period to come up with a reliable target value.

When calculating a mean, how many values do you recommend using?

AP: Usually, the laboratory information systems start calculating right away. You would do the second measurement, they would then come up with a mean, and then you can see it become stable. There are different philosophies. When I do this for an assay that I know is stable, I'm fine with a week.

Is there a troubleshooting action that I should consider as the most promising or most important?

AP: From my point of view, yes. The first instance should be to repeat the QC result to find out if you can reproduce the situation. If you can't reproduce it, maybe imprecision has increased, or you have occasional outliers that you might want to look at. Some laboratory information systems offer to take QC results out of the calculation.

However, I recommend you be honest and leave it in the calculation. When you release the patient result with an IQC, leave it in. Only take it out when you know you haven't done a patient result with it. So yes, from my point of view, repeating a failed QC is the most important troubleshooting action that you should always do. This should not keep you from doing other actions, even if the second IQC after that is good.

What is the difference between calibrators and quality controls?

AP: The main difference is the purpose. When you have a reagent, usually the manufacturer supplies the calibrators along with it. Whereas, when you have quality controls, ideally, they should come from a different vendor. They might be similar materials, or they might be different materials, the most important thing is that they are not the exact same material. Of course, it can be tough to produce quality controls and calibrators because often human material (i.e., human plasma or urine) or close-to-human material is needed. Then, you need to make sure that it's from different suppliers, that the target values assigned to it are from different methods, and that they're not out of the same system.

In 24/7 patient care, batch analysis becomes rare. How does that impact my IQC strategy?

AP: Batch analysis has a beginning and an end, and 24/7 patient care doesn't have that. So, we need to take into account a few points: Are the minimum requirements for the frequency of my QC good enough? Is it able to keep the quality of patient care at the desired level? I might want to increase the frequency and move away or do a little more that is required to be able to release results that are suitable for high-level patient care.

SelectScience runs 10+ webinars a month, discover more of our upcoming webinars>>